Anthropic to introduce invisible watermarks to Claude-generated text

Anthropic has announced it will begin watermarking text generated by its AI models, including Claude, to comply with new European regulations. The leading AI company confirmed this change in a recent update to its support page.

The EU AI Act’s Transparency Code, which came into effect on August 2, now requires AI companies to clearly mark content generated or edited by artificial intelligence so other systems can recognize it.

According to Anthropic, all models released after August 2 will automatically include technology to watermark both AI-generated text and files. For files, the company is relying on the C2PA open standard to embed these marks.

Anthropic also plans to extend watermarking support to older models. The company noted that the watermark will remain attached even when users copy and paste text elsewhere.

“Because the watermark is part of the text, it will travel with the text when it’s copied and pasted elsewhere, and may persist through some editing. Watermarking will be applied at the model level, which means it will be present no matter which Claude product or surface the text comes from,” the support page reads.

It remains unclear how much editing would be required to remove the watermark from AI-generated text. Anthropic has been asked for clarification, and any updates will be shared as they become available.

The company announced that watermarking will be implemented across a range of its products, including the Claude platform API, Claude, Claude Code, Claude Cowork, and CMany platforms are now moving quickly to add watermarks to AI-generated content, responding both to user concerns and increased regulatory attention.

READ MORE: Amazon Texas data center could become the largest US climate polluter

Just last week, the AI music platform Suno announced it would begin marking tracks created on its service following a series of legal challenges.

In June, newsletter platform Substack partnered with Pangram to help identify AI-generated content. Substack’s CEO, Chris Best, also highlighted the issue of ‘Claudefishing,’ a term used to describe people who use AI to produce content.

Apart from Anthropic, other companies such as Black Forest Labs, Google, Meta, Microsoft, OpenAI, and Synthesia have committed to complying with the EU’s code.